Privacy Policy

Arxivist · published by Cyquila

Effective: July 14, 2026  ·  Updated: October 5, 2026  ·  Contact: hello@cyquila.com

Summary

Arxivist is designed to be private by default. Your profile, bookmarks, and reading activity stay on your device, and paper ranking runs locally. We do not have accounts, and we never sell or share your personal information. To load papers quickly, the app asks a small paper-cache service run by Cyquila for each day’s public arXiv papers in the categories you follow — the section below describes exactly what that involves.

What the app stores (on your device only)

• A display name you optionally enter at first launch, saved locally and used only to personalize the interface.
• Your research profile — the arXiv categories you follow and the interest data computed from your inputs (author name, keywords, or topic picks), stored locally.
• Cached papers and bookmarks — arXiv paper metadata you’ve browsed or saved, kept on-device so the app works offline. Non-bookmarked papers are automatically removed after 14 days.

None of the above is ever sent to us or to anyone else.

What the app sends over the network

Arxivist fetches public arXiv papers from these sources, in this order:

1. The Arxivist paper-cache service (arxivist-api.cyquila.com) — a small, stateless service run by Cyquila on Cloudflare Workers that keeps a cached copy of each day’s public arXiv announcements so the app loads quickly. The app sends the arXiv category codes you follow (for example cs.LG) and a date, and receives that day’s public paper metadata. The request contains no name, profile, interest data, bookmarks, account, or device identifier. The service stores only copies of public arXiv paper data and an anonymous list of which category codes have been requested (used to keep popular categories fresh); it keeps no per-user records and no request logs. As with any internet service, the request travels through Cloudflare’s network, which processes technical data such as your IP address in order to deliver it — see Cloudflare’s privacy policy.
2. arXiv directly, if the cache service is unreachable — oaipmh.arxiv.org (a day’s announcement list) and rss.arxiv.org (the latest daily feed).
3. arxiv.org / export.arxiv.org — author publication lookups during onboarding, which always go to arXiv directly.

Requests to arXiv include only the category, author name, or date range needed to retrieve results, plus a standard app identifier. They do not include your name, profile, bookmarks, or anything traceable back to you. arXiv’s own privacy policy governs its services: info.arxiv.org/help/policies/privacy_policy.html.

On-device machine learning

Paper ranking runs entirely on-device using a bundled machine-learning model. Your interests and reading activity are never uploaded for training or any other purpose.

What we do not do

• No user accounts or sign-in required.
• No advertising, and no analytics or tracking SDKs.
• No collection of contacts, location, photos, or device identifiers.
• No selling or sharing of data (there is none to sell or share).

Children’s privacy

Arxivist does not knowingly collect personal information from anyone, including children under 13.

Your control

Your profile, bookmarks, and cached papers are stored only on your device, so you can remove everything by deleting the app, or by resetting your profile from the app’s Profile tab. The paper-cache service holds no data about you.

Affiliation

Arxivist is an independent application and is not affiliated with, endorsed by, or sponsored by arXiv or Cornell University. “arXiv” is a trademark of Cornell University.

Changes

We may update this policy as the app evolves. Material changes will be posted on this page with a new effective date.

Contact

Questions? hello@cyquila.com